Journal of Physics Conference Series

OPEN ACCESS

Fabrication of High Temperature ThermoelectricEnergy Harvesters for Wireless SensorsTo cite this article J E Koumlhler et al 2013 J Phys Conf Ser 476 012036

View the article online for updates and enhancements

You may also likeOdyssey of thermoelectric materialsfoundation of the complex structureKhalid Bin Masood Pushpendra Kumar RA Singh et al

-

Carrier mobilities thermoelectricproperties and band structures of type-Iclathrates Ba8M16Ge30 (M = Al Ga In)Kengo Kishimoto and Koji Akai

-

Energetic evaluation of possible interstitialcompound formation of BaSi2 with 2p- 3s- and 3d-elements using first-principlecalculationsYoji Imai Mitsugu Sohma and TakashiSuemasu

-

This content was downloaded from IP address 177154237254 on 10022022 at 2151

Fabrication of High Temperature Thermoelectric

Energy Harvesters for Wireless Sensors

JE Kohler1 R Heijl1 LGH Staaf1 S Zenkic2 E Svenman2AEC Palmqvist1 and P Enoksson1

1 Chalmers University of Technology Goteborg Sweden2 GKN Aerospace Trollhattan Sweden

E-mail elofstudentchalmersse

AbstractImplementing energy harvesters and wireless sensors in jet engines could simplify

development and decrease costs A thermoelectric energy harvester could be placed in thecooling channels where the temperature is between 500-900C This paper covers the synthesisof suitable materials and the design and fabrication of a thermoelectric module The materialchoices and other design variables were done from an analytic model by numerical analysisThe module was optimized for 600-800C with the materials Ba8Ga16Ge30 and La-dopedYb14MnSb11 both having the highest measured zT value in this region The design goalwas to be able to maintain a temperature gradient of at least 200C with high power outputThe La-doped Yb14MnSb11 was synthesized and its structure confirmed by x-ray diffractionMeasurement of properties of this material was not possible due to insufficient size of thecrystals Ba8Ga16Ge30 was synthesized and resulted in an approximated zT value of 083 at700C Calculations based on a module with 17 couples gave a power output of 1100mWg or600mWcm2 with a temperature gradient of 200K

1 IntroductionMeasuring pressure temperature stress and acceleration is essential for safe operation of jetengines For test engines in development this means thousands of sensors and kilometers of wireconnecting the sensors In the middle and the back of the jet engine it is more complicatedto place sensors due to both high temperature and the need for extensive cable wiring Cablewiring and the cost and weight of copper wire is a problem but can be reduced by replacing thewired sensors with wireless sensors coupled with energy harvesters as power sources

In the hot parts of the jet engine the walls are cooled down with air through cooling channelsPlacing thermoelectric energy harvesters inside the cooling channels would give a hot side of800-900C at the wall with cooling air of 500-600C on the cold side This means that thethermoelectric module could have a temperature gradient of 200C if the heat transfer throughthe module is low enough

This paper is a summary of an ongoing attempt to design and fabricate a small thermoelectricmodule optimized for the temperature range of 600-800C with high power output rather thanhigh efficiency and with capability to withstand a temperature gradient of 200C

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

Content from this work may be used under the terms of the Creative Commons Attribution 30 licence Any further distributionof this work must maintain attribution to the author(s) and the title of the work journal citation and DOI

Published under licence by IOP Publishing Ltd 1

2 DesignTo optimize the design for the temperature range and the environment an analytic model wasconstructed The model included heat input heat output Joule heating heat conduction [1]heat transfer in base plates parasitic thermal conductivity [2] and contact resistance [3] themodel also included load resistance matching the resistance of cables and DC-DC conversionlosses when connected to a voltage step up booster circuit The step up booster is used toincrease the voltage to an appropriate level to power the sensors in this case 33V

The choice of materials and variables such as number of couples in square devices heightof thermoelectric legs area of legs area ratio between n-type and p-type material thicknessof electrodes (resistance of contacts) load resistance base plates material electrode materialand material between couples (parasitic conductance) was done by numerical analysis of themodel These materials and variables were chosen to give the highest power output possiblewhen connected to a DCDC converter circuit The thermoelectric materials was chosen basedon simulations and the best results came from n-type Ba8Ga16Ge30 [4] and p-type La-dopedYb14MnSb11 [5] both with measured figure of merit above 12 in the temperature range Thesematerials also have similar thermal expansion with 142micromK [6] and 175micromK [7] respectively

3 SynthesisBa8Ga16Ge30 was synthesized by heating barium gallium and germanium with an 25 excessof barium to 1050C to ensure melting before cooled to 963C were it was held for 38h and thencooled at -100Ch to room temperature This synthesis method has been shown to producegood results [8] The result was 20g of large Ba8Ga16Ge30 single crystals with sizes up to 1cmin length see fig 1 The synthesis of La-doped Yb14MnSb11 followed a recipe [5] using a tinflux with an YbLaMnSbSn-ratio of 13161186 The crystals from this synthesis were smallerthan the Ba8Ga16Ge30 crystals and had tin impurities remaining on the crystals see fig 2

Figure 1 20g of Ba8Ga16Ge30 singlecrystals The size of the crystals wasrelatively consistent with some having alength of 1 cm making measurement ofSeebeck coefficient and resistivity possible

Figure 2 La-doped Yb14MnSb11 crystalscovered in a thin film of Sn No crystalswere big enough to enable measurement ofSeebeck coefficient electrical resistivity orthermal conductivity

4 FabricationThe module was built with alumina base plates which were cut into a size of 1cm2 Because La-doped Yb14MnSb11 reacts with most materials but is known to be stable with graphite alumina

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

2

and molybdenum [9] the choice of electrode material was molybdenum 80microm thick electrodeswere glued on to the base plates with high temperature ceramic glue The thermoelectric legswere cut with a diamond saw and placed in a grid of alumina paste which has been seen tosuppress sublimation of Yb14MnSb11 [10] The grid also stabilized the module and kept thethermoelectric legs in place during the fabrication process see fig 3

Figure 3 A grid is made out of an aluminapaste and placed on the base plates Thethermoelectric materials is placed in a gridwhich keeps the legs in place and stabilizesthe module when the top base plate has beenplaced

To ensure connection between the electrodes and the thermoelectric materials the electrodeswere shaped as springs Also a thin layer of graphite was placed between the electrodes andthermoelectric legs to improve the connection further The force needed to get a good connectionthrough the device was quite high and a large weight or a vise was necessary for low contactresistance Sealing the device was done with alumina paste and two different types of ceramicglues and was done to keep oxygen away from the thermoelectric materials and the electrodes

5 Results51 SimulationThe module design was based on the analytic model optimized for maximum power outputwith a temperature gradient of 200C Highest power output was given with 17-couples 1mmleg height and the optimal n-typep-type area ratio of 027073 Simulation of power outputvoltage and efficiency based on these values gave a power output of 750mW voltage of 640mVand an efficiency of 33 with a 200C temperature gradient of the environment see fig 4

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Temperature difference [K]

Pow

er [m

W]

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Vol

tage

[mV

] Figure 4 Estimated poweroutput (blue line) and volt-age (green dashed line) as afunction of the temperaturegradient of the environmentNo DCDC converter circuitis connected in this simula-tion

When including losses from cables and a low voltage DCDC converter with a load resistanceof 25Ω the power output is decreased to 123mW with a voltage of 1060mV In this simulationthe conversion efficiency in the step up converter was 30

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

3

05 1 15 2 25 3 35100

110

120

130

140

150

160

LoadVresistanceV[Ohm]

Pow

erV[m

W]

05 1 15 2 25 3 35600

700

800

900

1000

1100

1200

Vol

tage

V[mV

]

Figure 5 Estimated poweroutput (blue line) and volt-age (green dashed line) asa function of the load resis-tance The calculations arebased on 2m long high tem-perature cables commonlyused in jet engines with a wireresistance of 016Ωm and alow voltage step up converterwith 30 efficiency

52 SynthesisThe synthesis of Yb14MnSb11 resulted in crystals that were too small to do any measurementsof Seebeck coefficient resistivity or thermal conductivity The material was identified by x-raydiffraction measurements The synthesis of Ba8Ga16Ge30 gave single crystals large enough tomeasure Seebeck coefficient and resistivity but not thermal conductivity The measurement wasdone from room temperature to 725C and back to room temperature and showed negligiblehysteresis between the heat up and the cool down measurement The measured Seebeckcoefficient and resistivity during the cool down can be seen in fig 6 and 7

-2000

-1800

-1600

-1400

-1200

-1000

-800

-600

-400

0 100 200 300 400 500 600 700 800

S (micro

VK)

Temperature [oC]

Figure 6 Seebeck coefficient measurement on Ba8Ga16Ge30 The measurement was done from725C to room temperature No hysteresis could be seen between heat up and cool down

200

300

400

500

600

700

800

900

1000

0 100 200 300 400 500 600 700 800

ヾ ふび

cm)-

1

Temperature [oC]

Figure 7 Resistivity measurement on Ba8Ga16Ge30 Measured between 725C and roomtemperature while cooling down

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

4

000

010

020

030

040

050

060

070

080

090

0 100 200 300 400 500 600 700 800

zT

Temperature [oC]

Figure 8 Approximated figure of merit of Ba8Ga16Ge30 based on typical thermal conductivity[4] values from single crystals and the measured Seebeck coefficient and resistivity from thissynthesis

With no measurement of thermal conductivity the figure of merit of Ba8Ga16Ge30 could not becalculated However the difference in thermal conductivity for single crystal Ba8Ga16Ge30 issmall between different syntheses and an approximation based on known thermal conductivityresults [4] was done The approximation shows a figure of merit of 07-083 in the temperaturerange 600-725C see fig 8

6 Discussion and conclusionTo power a wireless sensor with a thermoelectric energy harvester the voltage should be stableand in the region of 25-35V The simulations when connected to a DCDC converter circuitshows a huge decrease in power output mainly because of the poor conversion efficiency fromthe low voltage and from the poor load resistance matching This problem can be improved byconnecting several thermoelectric modules in series to increased the voltage to make it possible touse more efficient DCDC converters Simulations with 4 modules in series give a power outputof approximately 2W with 33V output with 200C temperature gradient As most sensorsrequire only a few mW of power this energy harvester solution would give enough power for alarge sensor network

The approximated figure of merit of 083 at 700C for Ba8Ga16Ge30 is higher than what othercommon thermoelectric materials can achieve at this temperature When coupled with La-dopedYb14MnSb11 these materials could prove to be a powerful combination in the temperature rangeof 600-800C The successful synthesis of the n-type material and the steps so far tested showspromise for the complete assembly of a functional high temperature thermal harvester

References[1] Bitschi A 2009 Modelling of thermoelectric devices for electric power generation (Zurich Swiss Federal Institute

Of Technology Zurich) chapter 2[2] Omer S A and Infield D G 1998 Solar Energy Materials and Solar Cells 53 p 67-82[3] Rowe D M 2006 Thermoelectrics handbookmacro to nano (Boca Raton Taylor amp Francis Group) ch 39 p 1-15[4] Saramat A Svensson G Palmqvist A E C Stiewe C Mueller E Platzek D Williams S G K Rowe D M

Bryan J D and Stucky G D 2006 J Appl Phys 99 023708[5] Brown S Kauzlarich S Gascoin F and Snyder J 2008 Appl Phys Lett 93 062110[6] Okamoto N L Nakano T Tanaka K and Inui H 2008 J Appl Phys 104 013529[7] Ravi V Firdosy S Caillat T Brandon E Van Der Walde K Maricic L and Sayir A 2009 J Elec Mat 38

No 7 p 1433-42[8] Heijl R Cederkrantz D Nygren M Palmqvist A E C 2012 J Appl Phys 112(4) 044313[9] Paik J Brandon E Caillat T Ewell P and Feurial J Life testing of Yb14MnSb11 for high performance

thermoelectric couples 2011 Proceedings of Nuclear and Emerging Technologies for Space[10] Paik J and Caillat T 2010 NASA tech briefs aug p 22-23

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

5

2 DesignTo optimize the design for the temperature range and the environment an analytic model wasconstructed The model included heat input heat output Joule heating heat conduction [1]heat transfer in base plates parasitic thermal conductivity [2] and contact resistance [3] themodel also included load resistance matching the resistance of cables and DC-DC conversionlosses when connected to a voltage step up booster circuit The step up booster is used toincrease the voltage to an appropriate level to power the sensors in this case 33V

The choice of materials and variables such as number of couples in square devices heightof thermoelectric legs area of legs area ratio between n-type and p-type material thicknessof electrodes (resistance of contacts) load resistance base plates material electrode materialand material between couples (parasitic conductance) was done by numerical analysis of themodel These materials and variables were chosen to give the highest power output possiblewhen connected to a DCDC converter circuit The thermoelectric materials was chosen basedon simulations and the best results came from n-type Ba8Ga16Ge30 [4] and p-type La-dopedYb14MnSb11 [5] both with measured figure of merit above 12 in the temperature range Thesematerials also have similar thermal expansion with 142micromK [6] and 175micromK [7] respectively

3 SynthesisBa8Ga16Ge30 was synthesized by heating barium gallium and germanium with an 25 excessof barium to 1050C to ensure melting before cooled to 963C were it was held for 38h and thencooled at -100Ch to room temperature This synthesis method has been shown to producegood results [8] The result was 20g of large Ba8Ga16Ge30 single crystals with sizes up to 1cmin length see fig 1 The synthesis of La-doped Yb14MnSb11 followed a recipe [5] using a tinflux with an YbLaMnSbSn-ratio of 13161186 The crystals from this synthesis were smallerthan the Ba8Ga16Ge30 crystals and had tin impurities remaining on the crystals see fig 2

Figure 1 20g of Ba8Ga16Ge30 singlecrystals The size of the crystals wasrelatively consistent with some having alength of 1 cm making measurement ofSeebeck coefficient and resistivity possible

Figure 2 La-doped Yb14MnSb11 crystalscovered in a thin film of Sn No crystalswere big enough to enable measurement ofSeebeck coefficient electrical resistivity orthermal conductivity

4 FabricationThe module was built with alumina base plates which were cut into a size of 1cm2 Because La-doped Yb14MnSb11 reacts with most materials but is known to be stable with graphite alumina

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

2

and molybdenum [9] the choice of electrode material was molybdenum 80microm thick electrodeswere glued on to the base plates with high temperature ceramic glue The thermoelectric legswere cut with a diamond saw and placed in a grid of alumina paste which has been seen tosuppress sublimation of Yb14MnSb11 [10] The grid also stabilized the module and kept thethermoelectric legs in place during the fabrication process see fig 3

Figure 3 A grid is made out of an aluminapaste and placed on the base plates Thethermoelectric materials is placed in a gridwhich keeps the legs in place and stabilizesthe module when the top base plate has beenplaced

To ensure connection between the electrodes and the thermoelectric materials the electrodeswere shaped as springs Also a thin layer of graphite was placed between the electrodes andthermoelectric legs to improve the connection further The force needed to get a good connectionthrough the device was quite high and a large weight or a vise was necessary for low contactresistance Sealing the device was done with alumina paste and two different types of ceramicglues and was done to keep oxygen away from the thermoelectric materials and the electrodes

5 Results51 SimulationThe module design was based on the analytic model optimized for maximum power outputwith a temperature gradient of 200C Highest power output was given with 17-couples 1mmleg height and the optimal n-typep-type area ratio of 027073 Simulation of power outputvoltage and efficiency based on these values gave a power output of 750mW voltage of 640mVand an efficiency of 33 with a 200C temperature gradient of the environment see fig 4

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Temperature difference [K]

Pow

er [m

W]

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Vol

tage

[mV

] Figure 4 Estimated poweroutput (blue line) and volt-age (green dashed line) as afunction of the temperaturegradient of the environmentNo DCDC converter circuitis connected in this simula-tion

When including losses from cables and a low voltage DCDC converter with a load resistanceof 25Ω the power output is decreased to 123mW with a voltage of 1060mV In this simulationthe conversion efficiency in the step up converter was 30

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

3

05 1 15 2 25 3 35100

110

120

130

140

150

160

LoadVresistanceV[Ohm]

Pow

erV[m

W]

05 1 15 2 25 3 35600

700

800

900

1000

1100

1200

Vol

tage

V[mV

]

Figure 5 Estimated poweroutput (blue line) and volt-age (green dashed line) asa function of the load resis-tance The calculations arebased on 2m long high tem-perature cables commonlyused in jet engines with a wireresistance of 016Ωm and alow voltage step up converterwith 30 efficiency

52 SynthesisThe synthesis of Yb14MnSb11 resulted in crystals that were too small to do any measurementsof Seebeck coefficient resistivity or thermal conductivity The material was identified by x-raydiffraction measurements The synthesis of Ba8Ga16Ge30 gave single crystals large enough tomeasure Seebeck coefficient and resistivity but not thermal conductivity The measurement wasdone from room temperature to 725C and back to room temperature and showed negligiblehysteresis between the heat up and the cool down measurement The measured Seebeckcoefficient and resistivity during the cool down can be seen in fig 6 and 7

-2000

-1800

-1600

-1400

-1200

-1000

-800

-600

-400

0 100 200 300 400 500 600 700 800

S (micro

VK)

Temperature [oC]

Figure 6 Seebeck coefficient measurement on Ba8Ga16Ge30 The measurement was done from725C to room temperature No hysteresis could be seen between heat up and cool down

200

300

400

500

600

700

800

900

1000

0 100 200 300 400 500 600 700 800

ヾ ふび

cm)-

1

Temperature [oC]

Figure 7 Resistivity measurement on Ba8Ga16Ge30 Measured between 725C and roomtemperature while cooling down

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

4

000

010

020

030

040

050

060

070

080

090

0 100 200 300 400 500 600 700 800

zT

Temperature [oC]

Figure 8 Approximated figure of merit of Ba8Ga16Ge30 based on typical thermal conductivity[4] values from single crystals and the measured Seebeck coefficient and resistivity from thissynthesis

With no measurement of thermal conductivity the figure of merit of Ba8Ga16Ge30 could not becalculated However the difference in thermal conductivity for single crystal Ba8Ga16Ge30 issmall between different syntheses and an approximation based on known thermal conductivityresults [4] was done The approximation shows a figure of merit of 07-083 in the temperaturerange 600-725C see fig 8

6 Discussion and conclusionTo power a wireless sensor with a thermoelectric energy harvester the voltage should be stableand in the region of 25-35V The simulations when connected to a DCDC converter circuitshows a huge decrease in power output mainly because of the poor conversion efficiency fromthe low voltage and from the poor load resistance matching This problem can be improved byconnecting several thermoelectric modules in series to increased the voltage to make it possible touse more efficient DCDC converters Simulations with 4 modules in series give a power outputof approximately 2W with 33V output with 200C temperature gradient As most sensorsrequire only a few mW of power this energy harvester solution would give enough power for alarge sensor network

The approximated figure of merit of 083 at 700C for Ba8Ga16Ge30 is higher than what othercommon thermoelectric materials can achieve at this temperature When coupled with La-dopedYb14MnSb11 these materials could prove to be a powerful combination in the temperature rangeof 600-800C The successful synthesis of the n-type material and the steps so far tested showspromise for the complete assembly of a functional high temperature thermal harvester

References[1] Bitschi A 2009 Modelling of thermoelectric devices for electric power generation (Zurich Swiss Federal Institute

Of Technology Zurich) chapter 2[2] Omer S A and Infield D G 1998 Solar Energy Materials and Solar Cells 53 p 67-82[3] Rowe D M 2006 Thermoelectrics handbookmacro to nano (Boca Raton Taylor amp Francis Group) ch 39 p 1-15[4] Saramat A Svensson G Palmqvist A E C Stiewe C Mueller E Platzek D Williams S G K Rowe D M

Bryan J D and Stucky G D 2006 J Appl Phys 99 023708[5] Brown S Kauzlarich S Gascoin F and Snyder J 2008 Appl Phys Lett 93 062110[6] Okamoto N L Nakano T Tanaka K and Inui H 2008 J Appl Phys 104 013529[7] Ravi V Firdosy S Caillat T Brandon E Van Der Walde K Maricic L and Sayir A 2009 J Elec Mat 38

No 7 p 1433-42[8] Heijl R Cederkrantz D Nygren M Palmqvist A E C 2012 J Appl Phys 112(4) 044313[9] Paik J Brandon E Caillat T Ewell P and Feurial J Life testing of Yb14MnSb11 for high performance

thermoelectric couples 2011 Proceedings of Nuclear and Emerging Technologies for Space[10] Paik J and Caillat T 2010 NASA tech briefs aug p 22-23

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

5

and molybdenum [9] the choice of electrode material was molybdenum 80microm thick electrodeswere glued on to the base plates with high temperature ceramic glue The thermoelectric legswere cut with a diamond saw and placed in a grid of alumina paste which has been seen tosuppress sublimation of Yb14MnSb11 [10] The grid also stabilized the module and kept thethermoelectric legs in place during the fabrication process see fig 3

Figure 3 A grid is made out of an aluminapaste and placed on the base plates Thethermoelectric materials is placed in a gridwhich keeps the legs in place and stabilizesthe module when the top base plate has beenplaced

To ensure connection between the electrodes and the thermoelectric materials the electrodeswere shaped as springs Also a thin layer of graphite was placed between the electrodes andthermoelectric legs to improve the connection further The force needed to get a good connectionthrough the device was quite high and a large weight or a vise was necessary for low contactresistance Sealing the device was done with alumina paste and two different types of ceramicglues and was done to keep oxygen away from the thermoelectric materials and the electrodes

5 Results51 SimulationThe module design was based on the analytic model optimized for maximum power outputwith a temperature gradient of 200C Highest power output was given with 17-couples 1mmleg height and the optimal n-typep-type area ratio of 027073 Simulation of power outputvoltage and efficiency based on these values gave a power output of 750mW voltage of 640mVand an efficiency of 33 with a 200C temperature gradient of the environment see fig 4

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Temperature difference [K]

Pow

er [m

W]

0 20 40 60 80 100 120 140 160 180 2000

200

400

600

800

Vol

tage

[mV

] Figure 4 Estimated poweroutput (blue line) and volt-age (green dashed line) as afunction of the temperaturegradient of the environmentNo DCDC converter circuitis connected in this simula-tion

When including losses from cables and a low voltage DCDC converter with a load resistanceof 25Ω the power output is decreased to 123mW with a voltage of 1060mV In this simulationthe conversion efficiency in the step up converter was 30

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

3

05 1 15 2 25 3 35100

110

120

130

140

150

160

LoadVresistanceV[Ohm]

Pow

erV[m

W]

05 1 15 2 25 3 35600

700

800

900

1000

1100

1200

Vol

tage

V[mV

]

Figure 5 Estimated poweroutput (blue line) and volt-age (green dashed line) asa function of the load resis-tance The calculations arebased on 2m long high tem-perature cables commonlyused in jet engines with a wireresistance of 016Ωm and alow voltage step up converterwith 30 efficiency

52 SynthesisThe synthesis of Yb14MnSb11 resulted in crystals that were too small to do any measurementsof Seebeck coefficient resistivity or thermal conductivity The material was identified by x-raydiffraction measurements The synthesis of Ba8Ga16Ge30 gave single crystals large enough tomeasure Seebeck coefficient and resistivity but not thermal conductivity The measurement wasdone from room temperature to 725C and back to room temperature and showed negligiblehysteresis between the heat up and the cool down measurement The measured Seebeckcoefficient and resistivity during the cool down can be seen in fig 6 and 7

-2000

-1800

-1600

-1400

-1200

-1000

-800

-600

-400

0 100 200 300 400 500 600 700 800

S (micro

VK)

Temperature [oC]

Figure 6 Seebeck coefficient measurement on Ba8Ga16Ge30 The measurement was done from725C to room temperature No hysteresis could be seen between heat up and cool down

200

300

400

500

600

700

800

900

1000

0 100 200 300 400 500 600 700 800

ヾ ふび

cm)-

1

Temperature [oC]

Figure 7 Resistivity measurement on Ba8Ga16Ge30 Measured between 725C and roomtemperature while cooling down

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

4

000

010

020

030

040

050

060

070

080

090

0 100 200 300 400 500 600 700 800

zT

Temperature [oC]

Figure 8 Approximated figure of merit of Ba8Ga16Ge30 based on typical thermal conductivity[4] values from single crystals and the measured Seebeck coefficient and resistivity from thissynthesis

With no measurement of thermal conductivity the figure of merit of Ba8Ga16Ge30 could not becalculated However the difference in thermal conductivity for single crystal Ba8Ga16Ge30 issmall between different syntheses and an approximation based on known thermal conductivityresults [4] was done The approximation shows a figure of merit of 07-083 in the temperaturerange 600-725C see fig 8

6 Discussion and conclusionTo power a wireless sensor with a thermoelectric energy harvester the voltage should be stableand in the region of 25-35V The simulations when connected to a DCDC converter circuitshows a huge decrease in power output mainly because of the poor conversion efficiency fromthe low voltage and from the poor load resistance matching This problem can be improved byconnecting several thermoelectric modules in series to increased the voltage to make it possible touse more efficient DCDC converters Simulations with 4 modules in series give a power outputof approximately 2W with 33V output with 200C temperature gradient As most sensorsrequire only a few mW of power this energy harvester solution would give enough power for alarge sensor network

The approximated figure of merit of 083 at 700C for Ba8Ga16Ge30 is higher than what othercommon thermoelectric materials can achieve at this temperature When coupled with La-dopedYb14MnSb11 these materials could prove to be a powerful combination in the temperature rangeof 600-800C The successful synthesis of the n-type material and the steps so far tested showspromise for the complete assembly of a functional high temperature thermal harvester

References[1] Bitschi A 2009 Modelling of thermoelectric devices for electric power generation (Zurich Swiss Federal Institute

Of Technology Zurich) chapter 2[2] Omer S A and Infield D G 1998 Solar Energy Materials and Solar Cells 53 p 67-82[3] Rowe D M 2006 Thermoelectrics handbookmacro to nano (Boca Raton Taylor amp Francis Group) ch 39 p 1-15[4] Saramat A Svensson G Palmqvist A E C Stiewe C Mueller E Platzek D Williams S G K Rowe D M

Bryan J D and Stucky G D 2006 J Appl Phys 99 023708[5] Brown S Kauzlarich S Gascoin F and Snyder J 2008 Appl Phys Lett 93 062110[6] Okamoto N L Nakano T Tanaka K and Inui H 2008 J Appl Phys 104 013529[7] Ravi V Firdosy S Caillat T Brandon E Van Der Walde K Maricic L and Sayir A 2009 J Elec Mat 38

No 7 p 1433-42[8] Heijl R Cederkrantz D Nygren M Palmqvist A E C 2012 J Appl Phys 112(4) 044313[9] Paik J Brandon E Caillat T Ewell P and Feurial J Life testing of Yb14MnSb11 for high performance

thermoelectric couples 2011 Proceedings of Nuclear and Emerging Technologies for Space[10] Paik J and Caillat T 2010 NASA tech briefs aug p 22-23

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

5

05 1 15 2 25 3 35100

110

120

130

140

150

160

LoadVresistanceV[Ohm]

Pow

erV[m

W]

05 1 15 2 25 3 35600

700

800

900

1000

1100

1200

Vol

tage

V[mV

]

Figure 5 Estimated poweroutput (blue line) and volt-age (green dashed line) asa function of the load resis-tance The calculations arebased on 2m long high tem-perature cables commonlyused in jet engines with a wireresistance of 016Ωm and alow voltage step up converterwith 30 efficiency

52 SynthesisThe synthesis of Yb14MnSb11 resulted in crystals that were too small to do any measurementsof Seebeck coefficient resistivity or thermal conductivity The material was identified by x-raydiffraction measurements The synthesis of Ba8Ga16Ge30 gave single crystals large enough tomeasure Seebeck coefficient and resistivity but not thermal conductivity The measurement wasdone from room temperature to 725C and back to room temperature and showed negligiblehysteresis between the heat up and the cool down measurement The measured Seebeckcoefficient and resistivity during the cool down can be seen in fig 6 and 7

-2000

-1800

-1600

-1400

-1200

-1000

-800

-600

-400

0 100 200 300 400 500 600 700 800

S (micro

VK)

Temperature [oC]

Figure 6 Seebeck coefficient measurement on Ba8Ga16Ge30 The measurement was done from725C to room temperature No hysteresis could be seen between heat up and cool down

200

300

400

500

600

700

800

900

1000

0 100 200 300 400 500 600 700 800

ヾ ふび

cm)-

1

Temperature [oC]

Figure 7 Resistivity measurement on Ba8Ga16Ge30 Measured between 725C and roomtemperature while cooling down

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

4

000

010

020

030

040

050

060

070

080

090

0 100 200 300 400 500 600 700 800

zT

Temperature [oC]

Figure 8 Approximated figure of merit of Ba8Ga16Ge30 based on typical thermal conductivity[4] values from single crystals and the measured Seebeck coefficient and resistivity from thissynthesis

With no measurement of thermal conductivity the figure of merit of Ba8Ga16Ge30 could not becalculated However the difference in thermal conductivity for single crystal Ba8Ga16Ge30 issmall between different syntheses and an approximation based on known thermal conductivityresults [4] was done The approximation shows a figure of merit of 07-083 in the temperaturerange 600-725C see fig 8

6 Discussion and conclusionTo power a wireless sensor with a thermoelectric energy harvester the voltage should be stableand in the region of 25-35V The simulations when connected to a DCDC converter circuitshows a huge decrease in power output mainly because of the poor conversion efficiency fromthe low voltage and from the poor load resistance matching This problem can be improved byconnecting several thermoelectric modules in series to increased the voltage to make it possible touse more efficient DCDC converters Simulations with 4 modules in series give a power outputof approximately 2W with 33V output with 200C temperature gradient As most sensorsrequire only a few mW of power this energy harvester solution would give enough power for alarge sensor network

The approximated figure of merit of 083 at 700C for Ba8Ga16Ge30 is higher than what othercommon thermoelectric materials can achieve at this temperature When coupled with La-dopedYb14MnSb11 these materials could prove to be a powerful combination in the temperature rangeof 600-800C The successful synthesis of the n-type material and the steps so far tested showspromise for the complete assembly of a functional high temperature thermal harvester

References[1] Bitschi A 2009 Modelling of thermoelectric devices for electric power generation (Zurich Swiss Federal Institute

Of Technology Zurich) chapter 2[2] Omer S A and Infield D G 1998 Solar Energy Materials and Solar Cells 53 p 67-82[3] Rowe D M 2006 Thermoelectrics handbookmacro to nano (Boca Raton Taylor amp Francis Group) ch 39 p 1-15[4] Saramat A Svensson G Palmqvist A E C Stiewe C Mueller E Platzek D Williams S G K Rowe D M

Bryan J D and Stucky G D 2006 J Appl Phys 99 023708[5] Brown S Kauzlarich S Gascoin F and Snyder J 2008 Appl Phys Lett 93 062110[6] Okamoto N L Nakano T Tanaka K and Inui H 2008 J Appl Phys 104 013529[7] Ravi V Firdosy S Caillat T Brandon E Van Der Walde K Maricic L and Sayir A 2009 J Elec Mat 38

No 7 p 1433-42[8] Heijl R Cederkrantz D Nygren M Palmqvist A E C 2012 J Appl Phys 112(4) 044313[9] Paik J Brandon E Caillat T Ewell P and Feurial J Life testing of Yb14MnSb11 for high performance

thermoelectric couples 2011 Proceedings of Nuclear and Emerging Technologies for Space[10] Paik J and Caillat T 2010 NASA tech briefs aug p 22-23

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

5

000

010

020

030

040

050

060

070

080

090

0 100 200 300 400 500 600 700 800

zT

Temperature [oC]

Figure 8 Approximated figure of merit of Ba8Ga16Ge30 based on typical thermal conductivity[4] values from single crystals and the measured Seebeck coefficient and resistivity from thissynthesis

With no measurement of thermal conductivity the figure of merit of Ba8Ga16Ge30 could not becalculated However the difference in thermal conductivity for single crystal Ba8Ga16Ge30 issmall between different syntheses and an approximation based on known thermal conductivityresults [4] was done The approximation shows a figure of merit of 07-083 in the temperaturerange 600-725C see fig 8

6 Discussion and conclusionTo power a wireless sensor with a thermoelectric energy harvester the voltage should be stableand in the region of 25-35V The simulations when connected to a DCDC converter circuitshows a huge decrease in power output mainly because of the poor conversion efficiency fromthe low voltage and from the poor load resistance matching This problem can be improved byconnecting several thermoelectric modules in series to increased the voltage to make it possible touse more efficient DCDC converters Simulations with 4 modules in series give a power outputof approximately 2W with 33V output with 200C temperature gradient As most sensorsrequire only a few mW of power this energy harvester solution would give enough power for alarge sensor network

The approximated figure of merit of 083 at 700C for Ba8Ga16Ge30 is higher than what othercommon thermoelectric materials can achieve at this temperature When coupled with La-dopedYb14MnSb11 these materials could prove to be a powerful combination in the temperature rangeof 600-800C The successful synthesis of the n-type material and the steps so far tested showspromise for the complete assembly of a functional high temperature thermal harvester

References[1] Bitschi A 2009 Modelling of thermoelectric devices for electric power generation (Zurich Swiss Federal Institute

Of Technology Zurich) chapter 2[2] Omer S A and Infield D G 1998 Solar Energy Materials and Solar Cells 53 p 67-82[3] Rowe D M 2006 Thermoelectrics handbookmacro to nano (Boca Raton Taylor amp Francis Group) ch 39 p 1-15[4] Saramat A Svensson G Palmqvist A E C Stiewe C Mueller E Platzek D Williams S G K Rowe D M

Bryan J D and Stucky G D 2006 J Appl Phys 99 023708[5] Brown S Kauzlarich S Gascoin F and Snyder J 2008 Appl Phys Lett 93 062110[6] Okamoto N L Nakano T Tanaka K and Inui H 2008 J Appl Phys 104 013529[7] Ravi V Firdosy S Caillat T Brandon E Van Der Walde K Maricic L and Sayir A 2009 J Elec Mat 38

No 7 p 1433-42[8] Heijl R Cederkrantz D Nygren M Palmqvist A E C 2012 J Appl Phys 112(4) 044313[9] Paik J Brandon E Caillat T Ewell P and Feurial J Life testing of Yb14MnSb11 for high performance

thermoelectric couples 2011 Proceedings of Nuclear and Emerging Technologies for Space[10] Paik J and Caillat T 2010 NASA tech briefs aug p 22-23

PowerMEMS 2013 IOP PublishingJournal of Physics Conference Series 476 (2013) 012036 doi1010881742-65964761012036

5